The Real-time Multi-objects Detection & Tracking solution is developped on a novel machine learning base. The Lab uses machine learning algorithms and cascading architectures to detect vehicles in complex dynamic backgrounds. There are many scenarios, such as detecting on freeways, subways, suburban areas, and regular roads in several cities here in Taiwan.
After the intensive experimental tests, the detection accuracy is result to an average 93%, while the rate is at 85% if tested at spots with complex environments/backgroud. All these tests has proven the software solution can precisely detect vehicles and perform LPR (License Plate Recognition) at different distances on high-speed way (freeway).
NTUST also acts as the regional Industry-University Cooperation Center that has 21 TVE (Technical Vocational Education) schools in league. With all TVE schools' development force and resources, the center can coordinate and help enterprises in different field to develop new products from Taiwan with cutting-edge technologies and trends from time to time.